A running record of what we have shipped, hardened, and filed.
FLAG rules can now require human approval before an agent action executes. Actions are held in a pending queue. Named reviewers approve or reject with their email recorded against each decision. Python SDK v0.2.3 supports automatic polling via wait_for_approval().
Every guardrail rule change is now logged with the admin email, timestamp, and full before and after state. SHA-256 hash chaining ensures the log is tamper-evident. Exportable as CSV.
Six new rule types added alongside the existing PII Shield and Spend Limit: Keyword Block, Keyword Flag, Rate Limiter, API Scope Control, Regex Pattern Match, Output Length Limit, Prompt Injection Detector, and Repetition Detector.
Three pre-configured compliance packs added: EU AI Act Article 12 Pack, HIPAA Safety Pack, and Financial Services Pack. Each pack adds multiple pre-configured rules with one click.
Free 30-day silent observation of your AI agent activity. No code changes. No commitment. Request at gateplex.ai/shadow-audit.
Gateplex VPC Edition is now available. The governance engine runs entirely inside your AWS or Azure environment via a four-container Docker stack. Zero data egress. Air-gapped deployment available on request.
Second provisional patent application filed covering the VPC split-plane deployment architecture and per-record SHA-256 hash-chained compliance evidence.
Official Python SDK published with OpenAI and Anthropic one-line patchers, batch intercept, context manager, and PENDING_APPROVAL verdict support.
Gateplex MCP server listed on Smithery, Glama, and MCP.so. Discoverable by any MCP-compatible AI agent framework.
First provisional patent application filed covering the core interception architecture and tamper-evident audit trail mechanism.
Gateplex governance engine live at gateplex.ai. Server-side enforcement, per-record SHA-256 hash chaining, live intercept feed, compliance PDF export.